CT scan changes over one year predict outcomes in fibrotic lung disease

Researchers at National Jewish Health have shown that subtle increases in lung scarring, detected by an artificial intelligence-based tool on CT scans taken one year apart, are associated with disease progression and survival in patients with fibrotic interstitial lung disease. The findings, recently published in the American Journal of Respiratory and Critical Care Medicine, suggest that computer-based image analysis may provide an earlier, more objective way to identify patients at highest risk for worsening disease.

“We found that even small increases in fibrosis over one year signal a higher risk of lung function decline and mortality,” said Matthew Koslow, MD, pulmonologist at National Jewish Health and lead co-author of the study. “For example, patients with a 5% or more increase in fibrosis score showed a greater than two-fold increased risk of death or lung transplant and steeper declines in lung function in the following year compared to patients with stable fibrosis scores. What is especially important is that these changes were strongest in patients with less severe disease at baseline —precisely the group where earlier intervention has the greatest potential to alter the course of disease.”

Fibrotic interstitial lung diseases, which include idiopathic pulmonary fibrosis, or IPF, are a group of chronic, progressive disorders marked by lung scarring that make breathing increasingly difficult. Current tools for predicting progression rely on symptoms, lung function tests and radiologist interpretation of high-resolution CT scans—each of which can be limited by subjectivity or variability, especially when evaluating changes over time.

The new study used a deep learning method called data-driven textural analysis (DTA). Developed by the Quantitative Imaging Laboratory at National Jewish Health, DTA provides a precise measurement of the extent of lung fibrosis on CT scans. Researchers found that increases in DTA fibrosis scores over one year were strongly associated with subsequent lung function decline and a higher risk of death or lung transplant.

“This work demonstrates how quantitative imaging and robust statistical modeling can uncover meaningful patterns in disease progression,” said David Baraghoshi, Ph.D., biostatistician at National Jewish Health and co-first author of the study. “By analyzing changes in fibrosis scores over time and linking them to future outcomes, we were able to show that imaging data can serve as a powerful marker of clinical trajectory.”

The results were validated using data from the Pulmonary Fibrosis Foundation Patient Registry, underscoring the generalizability of the findings.

These insights could have major implications for clinical trials and patient care. Quantitative CT analysis may serve as a meaningful trial endpoint, a tool for selecting patients at highest risk, and a guide for treatment decisions in real-world practice.